TVPI, DPI and IRR

Endowments and pensions allocate to venture because they want returns that beat public markets enough to justify illiquidity. A rough planning target many GPs talk about is on the order of ~3× net to LPs over the fund life (sometimes framed as mid-teens net IRR in a good vintage). That bar, after fees and carry, is why a book of solid 2× outcomes does not count as success.

The big idea

LPs eventually need cash they can spend on pensions and university budgets. Marks on unsold companies feel like wealth. They are not. The three numbers below separate the two.

NAV (net asset value) is the marked value of unrealized holdings — often based on the last round price. In boom years TVPI looks heroic while DPI stays low. LPs eventually care about DPI. A fund that marks unicorns but never distributes cash cannot pay pensions.

IRR is sensitive to when cash goes out and comes back. Early exits help IRR even if the multiple is the same.

Walking through a 3× LP target

For a $100M fund to deliver ~$300M net to LPs, the portfolio must produce well over $300M gross. Fees have already consumed ~$20M of buying power. Carry takes 20% of profits. Working backwards: if LPs need $300M out, and carry is 20% of profit after capital return:

LPs receive capital back ($100M) + 80% of profit
Want LP total ≈ $300M
⇒ profit to share ≈ ($300M − $100M) / 0.8 = $250M
⇒ gross distributions ≈ $100M + $250M = $350M

In plain language: the companies must return on the order of 3.5×+ on committed capital, and closer to 4×+ on the ~$80M actually invested, for LPs to see a clean 3×. Partners do not celebrate a fund that "kind of worked."

This is why the power law dominates strategy. One or two fund-returning companies do the job. A stack of modest wins does not.

Common mistakes

• Reading a high TVPI as a done fund. Unrealized NAV can unwind in the next down round.
• Ignoring timing. Two funds with the same multiple can have very different IRRs.
• Treating ~3× net as optional. That is the planning bar many LPs use to justify the asset class.

Check yourself

TVPI is 4× and DPI is 0.4×. What is the fund rich in?

Paper marks (NAV). Cash returned to LPs is still small.

Why does a 3× net-to-LP target imply more than 3× gross from the companies?

Fees reduced buying power, and carry takes 20% of profits before LPs see the rest.

Key takeaways

• A common LP planning target is on the order of ~3× net, which needs ~3.5×+ gross.
• TVPI includes paper value. DPI is cash. IRR cares about timing.
• LPs eventually demand DPI. Marks are not distributions.

Further learning

Next, reserves and ownership — how funds try to own enough of the winners to clear this bar.